[all-commits] [llvm/llvm-project] fe3933: [mlir][vector] Support complete folding in single ...

[Yang Bai via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: fe3933da15b5bc635bce156f1f8d11a784316a07
      https://github.com/llvm/llvm-project/commit/fe3933da15b5bc635bce156f1f8d11a784316a07
  Author: Yang Bai <baiyang0132 at gmail.com>
  Date:   2025-06-18 (Wed, 18 Jun 2025)

  Changed paths:
    M mlir/lib/Dialect/Vector/IR/VectorOps.cpp
    M mlir/test/Dialect/Affine/constant-fold.mlir
    M mlir/test/Dialect/Linalg/mesh-spmdization.mlir
    M mlir/test/Dialect/Mesh/spmdization.mlir
    M mlir/test/Dialect/Tensor/mesh-spmdization.mlir
    M mlir/test/Dialect/Tosa/constant_folding.mlir
    M mlir/test/Dialect/Vector/constant-fold.mlir
    A mlir/test/Dialect/Vector/single-fold.mlir
    M mlir/test/Transforms/constant-fold-debuginfo.mlir
    M mlir/test/Transforms/constant-fold.mlir
    M mlir/test/lib/Transforms/CMakeLists.txt
    R mlir/test/lib/Transforms/TestConstantFold.cpp
    A mlir/test/lib/Transforms/TestSingleFold.cpp
    M mlir/tools/mlir-opt/mlir-opt.cpp

  Log Message:
  -----------
  [mlir][vector] Support complete folding in single pass for vector.insert/vector.extract (#142124)

### Description

This patch improves the folding efficiency of `vector.insert` and
`vector.extract` operations by not returning early after successfully
converting dynamic indices to static indices.

This PR also renames the test pass `TestConstantFold` to
`TestSingleFold` and adds comprehensive documentation explaining the
single-pass folding behavior.

### Motivation

Since the `OpBuilder::createOrFold` function only calls `fold` **once**,
the current `fold` methods of `vector.insert` and `vector.extract` may
leave the op in a state that can be folded further. For example,
consider the following un-folded IR:
```
%v1 = vector.insert %e1, %v0 [0] : f32 into vector<128xf32>
%c0 = arith.constant 0 : index
%e2 = vector.extract %v1[%c0] : f32 from vector<128xf32>
```
If we use `createOrFold` to create the `vector.extract` op, then the
result will be:
```
%v1 = vector.insert %e1, %v0 [127] : f32 into vector<128xf32>
%e2 = vector.extract %v1[0] : f32 from vector<128xf32>
```
But this is not the optimal result. `createOrFold` should have returned
`%e1`.
The reason is that the execution of fold returns immediately after
`extractInsertFoldConstantOp`, causing subsequent folding logics to be
skipped.

---------

Co-authored-by: Yang Bai <yangb at nvidia.com>



To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications